Explain this asphalt sheen

[.Scott]

Homework Statement

This isn't actually homework - but it's like homework and I am not completely certain about the full answer. Most importantly, I believe it's entertaining.

The statement is this: They have just repaved the parking lot in back of my work place. As I look down on the fresh asphalt from a 2nd floor corridor, I see a sheen - wide bands of rainbow-like colors. Each of these bands is about a meter wide and a full cycle of colors is seen about every 4 meters. The corridor is lined with plate glass panes, each about 2 meters wide. As I walk down the corridor, the color pattern on the pavement stays fixed to positions on the pavement until I cross to the next pane. Then it shifts as I cross to that next pane. So, for example, from one pane, I see the purple ban lined up with a particular fire hydrant. As I continue walking across that pane, the purple ban remains in that exact position. But when cross the divider between panes, the view through the next pane shows the pattern shifted and now the hydrant is lined up with a yellow ban.
Explain this affect.

Relevant Equations

This is a word problem. I believe it can be answered without resorting to equations.

This problem is two parts. The first is to determine what effects are being provided by each of the elements - 1) the window panes; 2) the asphalt surface. My answer to that is

Spoiler

Each pane is polarizing light at a somewhat different angle. And: As you move across the asphalt surface, the oil thickness will change - and the resulting constructive and destructive interference will affect different color wavelengths differently.

The second part of the problem is exactly why you get this affect.

Spoiler

Clearly, the oily layer affects both the interference and polarization based on the wavelength. But I imagine there must be some easy way of describing why the polarization is not only color-dependent, but also has the affect of rotating the color phases of the sheen.

And one more spoiler:

Spoiler

One of the panes does not show the sheen at all. Or, at least, not that I could notice. Compared to the others, its like looking at a black and white photo.

 

[berkeman]

No pictures?

 

[.Scott]

No pictures?

I don't bring my cell phone to work.

 

[berkeman]

I don't bring my cell phone to work.

Okay, fair enough. Since that may involve workplace security considerations, I won't say anything further on that.

I basically need to carry a cellphone 24/7, due to my work as a medic and also due some volunteer channels that I participate in like PulsePoint (nearby CPR emergency notifications).

 

[.Scott]

It's still there today - although a bit faded.
On closer inspection, it's looking more like a rainbow than ever. The number of bands is limited. The colors are: a stubby yellow band, a minimal purple band, then blue, green, yellow, purple, blue, green, yellow, purple blue, and a minimal green. To the left and right of this pattern, there is no sheen. But the bands stretch out across the parking lot to its distant edge. So - a triple rainbow.

I found an image showing the same effect in Reddit:

Here is a link that mentions the polarization of light in the kind of meteorlogical rainbows that we are all familiar with. According to that posting, water rainbow light can be 94% polarized.

However, what I am seeing (and what is shown in the photograph) is not a water-based rainbow. the conditions outside have been warm and sunny - plenty enough to dry out any water. And the photo above shows the sheen in direct sub (note shadows of power lines).

Also, this is not the result of thin-film-interference. Notice that in the photo above, there is a hint of the color bands even in the concrete in the foreground.

And in my view outside, I also see the pattern resume from one side of a median strip to the other.

My current best guess is that this is an effect of silicon-dioxide.

 

[DaveC426913]

I have never seen anything remotely like that effect.

 

[berkeman]

Did it say in the Reddit post what kind of vehicle that is and what kind of window treatment it has?

Also, if you can still see the effect at work, can you try a pair of polarized sunglasses to see if rotating the sunglasses changes the effect?

 

[.Scott]

I will not be at that work location tomorrow and I expect that on Monday that lot will have the lines painted and be in use (including my use).
I will see if I can find an old pair of Polaroids.
Of course, the glasses will be of little if any use when viewed through a polarizing window.

The photo above comes from this link. ...
... which is entitled "Polarized glasses creates effect of rainbow road".
It was posted 3 years ago - and several posters identified the effect as the stress marks you can see in some windows when wearing polarizing glasses. However, I think you can tell, that's not what it is.
However, I do believe that that side window is up. First because those windows don't usually roll completely down. Second, because that white "spray" looks like a reflection of the base of the windshield to me.

The color saturation in that photo is roughly double what I saw in the parking lot. I credit that to the "incidental polarization" of the building windows vs. the intentional polarization of the polarizing glasses.

And this color saturation is much better than the usual rainbow - likely because the fresh pavement provides a very dark background - in contrast to the fairly bright sky behind the usual rainbows.

 

[.Scott]

That new parking lot is now open - and I brought polarizing sunglasses - and the effect is still apparent.
From ground level, there is nothing to see. It is all black. On close inspection, it is not that smooth and there are no sand particles or anything else other than what appears to be black-covered gravel.

There were some heavy rains this weekend and some of the pavement is still wet. It's a bright morning sun coming in low.

Using (or not using) the polarizing sunglasses at ground level, the pavement appears black and reflects some glare from the sun - but no colors.

Using the polarizing sunglasses and looking down at the pavement from an unpolarized second story, the color pattern is evident. Rotating the polarization angle changes the color pattern. A 90-degree rotation results in all the reds and greens switching places. A 45-degree rotation results in something different - I will spent a bit of time on it later today and see if I can't describe it.

The colors, as viewed through the windows, are not as saturated as they were last week. The view from the sunglasses shows the colors very clearly - though not as well as that Reddit photo.

 

[.Scott]

I have more info:
Now that I have polarized glasses and there are cars in the lot, I can see something else: The windshields on many of the cars also show the color. And - the color is 100% position dependent. So if I look at a car parked in a blue band of the pavement, if it has a "participating windshield (PW)", it will reflect blue. If I rotate the glasses or look in some area with red pavement, PWs will appear red.

Also, in the same way that if you look at the reddit photo, you can see the banding continue into the pavement that has a much lighter color, I can see that same effect where the newly paved parking lot borders the older lot.

I said I was going to look more carefully at the rotating polarization and what you get at 45 degrees of rotation. It is as I was originally expecting, a smooth transition through the color spectrum. So, after 90 degrees, red and green switch places; but after 45 degrees red ends up in what was the red-to-green transition area and green ends up in what was the green-to-red transition area.

 

[.Scott]

So, given these new observations posted above (and some related to the details of the window polarization), here's my heavily revised hypothesis:

1) The sunlight starts out unpolarized and illuminated the entire parking lot.
2) Many surfaces, including the pavement itself, reflect horizontally polarized light more than vertically polarized light. So the light reaching the windows in the building is predominantly horizontally polarized.
3) The windows are slightly polarized horizontally, but more importantly, they exhibit "optical rotatory dispersion (ORD)" which is both wavelength and approach angle dependent. The wavelength dependency is due to the speed at which that wavelength travels through the transparent material. The approach angle dependency is due to the length of the path of the light through the transparent material.
4) Most windows affect the polarization in pretty much the same way. But a few appear different. I originally described them as unpolarized - but on further study, they simply have a different thickness of the ORD material (probably less).
5) So when the light exits the window pane and enters the corridor, it will still be polarized - but not usually horizontally polarized.
6) When observed casually, the incidental horizontal polarization of the pane will allow the observer to see the colorful effect.
7) When observed through sunglasses (vertically polarized) a very similar effect is seen - but the color pattern is 180 degrees out of phase - and the colors are for more brilliant.

As a bonus, the sunglasses allow me see color patterns in the clouds that cannot be seen without them.

This also suggests that the Reddit photo must have been taken through a similar ORD material. But when asked about car side-windows, Google's AI expertly stated to me that that was not possible because silicon dioxide is "inherently achiral".

So I did a more detailed Google search and it came up with this 11-year-old link:
Car Window Tinting (Physics Forum).

So it would seem that not only was the Reddit photo taken through a car window (which was as we expected), but that it was taken through a tinted window. And most certainly, the windows along the corridor are tinted.

One more note: This is basically color-encoded polarizations. When I web-searched for photographs of this chiral color effect, I got tons of microscope photos, but no outdoor scenes. The world is missing something big and colorful.